How can CORDEX enhance assessments of climate change impacts and adaptation?

Size: px

Start display at page:

Download "How can CORDEX enhance assessments of climate change impacts and adaptation?"

Ruth Simmons
5 years ago
Views:

1 How can CORDEX enhance assessments of climate change impacts and adaptation? Timothy Carter Finnish Environment Institute, SYKE Climate Change Programme

2 Outline Priorities for IAV research Demand for climate information Added value of downscaling Framing uncertainties Conclusions

3 What's this IAV anyway? Climate Change Impacts, Adaptation and Vulnerability but not everyone agrees...

4 The evolution of risk assessment as applied to climate change, particularly adaptation, since the establishment of the IPCC in 1988 Source: Jones and Preston (2011)

5 Outline Priorities for IAV research Demand for climate information Added value of downscaling Framing uncertainties Conclusions

6 Outline Priorities for IAV research Demand for climate information Added value of downscaling Framing uncertainties Conclusions

7 Some key motivations of users in requesting climate information Obtaining general information about climate change major messages at various scales Obtaining information tailored to user requirements and adaptation decision-making Communicating uncertainties in future projections Ensuring comparability across assessments for coordinating integrated responses Reconciling projections with recent trends and with planning/policy scenarios Addressing changes in frequency and magnitude of extreme weather events

8 Who are the IAV community? Empiricists Observed impacts (long-term/extreme events); observed adaptation (reactive and anticipatory; to averages/extremes; maladaptation) Impact modellers Disciplinary/process focus (and ESMs); disciplinary/applied focus; inter disciplinary/applied focus; integrated assessment; global to local Experimentalists Gas enrichment; controlled climate; materials; land use management Vulnerability assessors Mapping indices/indicators; local (household) to global Adaptation researchers Adaptation processes; adaptation practices; management for adaptation and mitigation; development of analytical methods, tools and metrics Adaptation policy analysts Governance; coherence; integration; mitigation/adaptation; emergency planning; implementation Well, that's just a sample!

9 How many of these can make use of downscaled climate projections? Empiricists Qualitatively; could help to frame studies (e.g. surveys, questionnaires) Impact modellers High demand, though not all merited; impacts of extreme weather; variability change; differential impacts over varied terrain Experimentalists Little use except perhaps qualitatively for experimental design Vulnerability assessors Can offer added value for quantifying certain indicators Adaptation researchers Informing about climate risk; identification and appraisal of adaptation measures Adaptation policy analysts Awareness raising of future regional climate, including extremes

10 Outline Priorities for IAV research Demand for climate information Added value of downscaling Framing uncertainties Conclusions

11 Outline Priorities for IAV research Demand for climate information Added value of downscaling Framing uncertainties Conclusions

12 What more does downscaling offer? Conditional on the global projection downscaled: Finer-scale spatial resolution More realistic physical representation of local climate Improved resolution of extreme weather events [Maybe] enhanced realism in estimates of future climate Note: Downscaling also throws up issues concerning the quality of "non-climate" information required at the same spatial scale (e.g. social, economic, land use, other environmental stresses)

13 Methods of regional climate change scenario construction, listed in the order of increasing complexity and resource demand (example adaptation activities in parentheses) # Source: Wilby et al. (2009)

14 Methods of regional climate change scenario construction, listed in the order of increasing complexity and resource demand (example adaptation activities in parentheses) # Source: Wilby et al. (2009)

15 Downscaling lessons for hydrology 1. What more (if anything) can be learnt from downscaling method comparison studies? Little more from intercomparison studies; recommend "sensitivity" step to determine key climate variables and most appropriate downscaling method CORDEX? 2. Can dynamical downscaling contribute advantages that can not be conferred by statistical downscaling? Yes. Regional climate change signals can differ from GCMs, particularly in regions with complex orography; improved simulation of higher moment climate statistics 3. Can realistic climate change scenarios be produced from dynamically downscaled output for periods outside the time period of simulation using methods such as pattern scaling? Yes for temperature; maybe for other variables; but transient runs now common 4. What new methods can be used together with downscaling to assess uncertainties in hydrological response? Probabilistic methods offer promise of more robust treatment of uncertainties 5. How can downscaling methods be better utilized within the hydrological impacts community? Comparison of offline or coupled online hydrological modelling (with feedbacks); testing if hydrological models are "fit for purpose" for assessing climate change impacts Fowler et al. (2007)

16 Outline Priorities for IAV research Demand for climate information Added value of downscaling Framing uncertainties Conclusions

17 Outline Priorities for IAV research Demand for climate information Added value of downscaling Framing uncertainties Conclusions

18 WCRP CORDEX Downscaling domains (Coordinated Regional Downscaling Experiment) Source: Colin Jones, 2010

20 What new long-term climate information does ENSEMBLES offer impact analysts? RCM and GCM projections (Giorgi regions) John Caesar (Met Office), Stefan Fronzek (SYKE), Ines Hoeschel (FUB), Philip Lorenz (MPI), Aurore Voldoire (CNRM)

21 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

22 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

23 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

24 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

25 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

26 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

27 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

28 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

29 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

30 What new long-term climate information does ENSEMBLES offer impact analysts? "Grand ensemble" PDFs (sampling multiple uncertainties) Glen Harris (Met Office), Stefan Fronzek (SYKE)

31 Decomposing the UKCP09 methodology used in creating probabilistic projections for Europe ("grand ensemble") Global analysis ("grand ensemble") Based on: Murphy et al. (2007)

32 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

33 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

34 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

35 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

36 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

37 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

38 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

39 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

40 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

41 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

42 What new long-term climate information does ENSEMBLES offer impact analysts? ENSEMBLES multiple projections intercomparison

43 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

44 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

45 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

46 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

47 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

48 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

49 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

50 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

51 Northern Europe Projected mean annual temperature ( C) and precipitation (%) change relative to Southern Europe

52 24 "Giorgi" regions as presented in Ch 11, WG I AR4 UK Met Office Hadley Centre: paper under revision on estimating "grand ensemble" uncertainties for Giorgi regions (Harris, pers. com.)

53 Role of impact model uncertainty Change in net farm revenues for Sri Lanka ($/Ha) under a range of climate model projections and emissions scenarios Impact model uncertainty Emissions uncertainty Climate model uncertainty Source: Wilby et al. (2009)

Changes in median flow simulated with a hydrological model (CATCHMOD) when model parameter uncertainties are combined with the climateprediction.net (CP.

54 Changes in median flow simulated with a hydrological model (CATCHMOD) when model parameter uncertainties are combined with the climateprediction.net (CP.net) ensemble of climate model projections Light Dark Red Green Black cross: curve: blue curves: curves: curve; curve: standard all possible each each default standard CATCHMOD HadSM3 climate combinations HadSM3 projection; climate version; climate version; projection; projection; CATCHMOD climate 449 default climate projections 100 versions CATCHMOD versions Source: New et al. (2007)

56 Winter wheat yields simulated (eight models) and observed at a site in the Czech republic, Mean (blue) is the distribution of the mean predictions from the eight models in different years Palosuo et al. (submitted)

57 Outline Priorities for IAV research Demand for climate information Added value of downscaling Framing uncertainties Conclusions

58 Outline Priorities for IAV research Demand for climate information Added value of downscaling Framing uncertainties Conclusions

59 Research imperatives regarding applications of regional climate research (early 2000s) Applications Improve availability and accessibility of regional climate information Improve realism of driving factors and representation of processes (e.g. biogeochemistry; lakes) Involve stakeholders in determining the required resolution of regional climate information Investigate other applications of regional climate information (e.g. storm surges; air quality) Overall Need for co-ordinated "end-to-end" prediction systems for impact assessment Seasonal prediction can inform added value of downscaling Funding agencies need to recognise importance of model development and evaluation in addition to prediction All downscaling techniques produce useful results continue parallel activities Leung et al. (2003)

61 Increasing complexity /resource demand Appropriateness of different scenario methods for representing climate of the 2020s in light of the varying needs of IAV analysts Inappropriate Neutral or depends Appropriate Indicator Preferred attributes for development and adaptation planning Capacity building Low personnel, technical and infrastructure requirements Low data, time and financial costs High spatial resolution (site or region, not continental or global) High temporal resolution (hourly or daily, not monthly or annual) High realism and joint behaviour of weather variables High ability to represent different external forcing (land cover, aerosols) High capability for providing probabilistic information High ability to produce surfaces or maps of climate change High ability to produce transient (rather than time-slice) scenarios High availability of tools, supporting data and guidance CORDEX Modified from Wilby et al. (2009)

62 Making CORDEX results effective Information generated in consultation with stakeholders Projections available in accessible formats Tools offered for extracting data to suit applications Regionalisation of information delivery (e.g. through portals) Observational climate data available at equivalent resolution to projections (including reanalysis) Documentation of the data and projections including quality information (e.g. missing data, errors, etc.) Summary information, including graphs, maps, statistics but also clear narratives Contextual regional information for framing uncertainty (e.g. from global models) Guidance on the use and misuse of downscaled information, including examples

63 Examples of international climate data and scenario web portals

68 Take home messages Use of downscaled projections in IAV assessment Don't: Do: Use single projections to inform decisions Overlook impact uncertainties Expect multiple projections to represent all uncertainties Assume that finer resolution = more accurate projections Consult stakeholders to prioritise projection needs Explore impacts of extreme weather Examine impacts of climate variability at different scales Compare impacts with those for bounding GCM(s) Where possible, apply a range of downscaling methods

69 Notice Colleagues are welcome to incorporate these slides into their own presentations, assuming they are correctly acknowledged. However, the author would also appreciate being informed prior to the extensive use of this material in public meetings.

ENSEMBLES. ENSEMBLES Final Symposium, November 2009, Met Office Exeter Page 1

ENSEMBLES. ENSEMBLES Final Symposium, November 2009, Met Office Exeter Page 1 ENSEMBLES ENSEMBLES Final Symposium, 17-19 November 2009, Met Office Exeter Page 1 Setting the scene for the third project aim: Maximising the results by linking the ensemble prediction system to a range